AdsenseLink

星期五, 3月 31, 2006

上週預測結果

上周嘗試預測了恆指和四隻股票, 其中恆生指數,SCMP(583), 和黃(13)如預期中下跌. 雖然中銀(2388)也下跌了, 但跌幅卻很輕微, 看來$15.5的價位, 支持力十足. 同仁堂(8069)出奇地上升了, 破壞了我預測模型的準確度. 但我還是感到好高興, 因為手持重貨, 只是可惜不能再買多一些平價貨. 不過話番轉頭, 同仁堂(8069)交投量太少, 容易被人為操縱, 也許並不適合用 Regression 的方法來預測.星期三調校了預測模型並用和黃來做測試, 準確度是好像有所提高. 將繼續測試和改進.可惜下星期秋官殺到, 此君令受股市增添變數, 更難預測. 所以下星期還是避一避風頭, 暫停預測的測試.但各位人兄有股票想參考小弟的預測模型, 可以在小弟的Blog 中留言, 如抽到時間, 我會製圖並Post 出. 但強調只供參考, 本人自己都持懷疑態度.

星期三, 3月 29, 2006

利用和黃調整預測參數

有時候不同的預測參數會會得出完全相反的預測結果, 要在實戰加以調整. 總覺得之前的預測結果強差人意, 所以根據舊數據作了一些測試, 發現 Stepwise Autoregression 加上 Quadratic Time Trend Model 的效果似乎比較好. 今日特登用和黃來測試, 看看結果如何.和黃(0008) 下星期似乎會繼續下跌直到 $70.5. 真奇怪, 最近長實升了不少, 不知為何和黃卻一直下跌. 照計之前李嘉誠都入左不少. 吾通靠害?

星期日, 3月 26, 2006

中銀香港走勢仍然下跌



受人所托, 對中銀香港作了一次預測, 結果有點出乎意料 , 中銀香港走勢向下,下星期有跌破$15的可以. 有於一直沒有研究此股, 所以也不肯定有幾準, 各位點睇?

星期六, 3月 25, 2006

恆指預測

只供?考. Heng Seng Index Reference

上週預測結果

上週嘗試預測了四隻股票, 成績還算可以. 其中和黃保持強勢, 沒有如預期中下跌, 我想主要是因為 Expotential Smoothing 的方法和每周數據不適合作為短期預測. 改為 Stepwise Regression 和每日數據後, 準確度好像有所提高. 尤其是魏橋紡織, 曾大幅下跌致 $11.75, 只是星期五回升了不少.壹傳媒 (283) 雖然如預期下跌, 但跌幅比較輕微




今個主要觀察兩隻股票, 同仁堂(8069)和南華早報 (583) 結果如下.




同仁堂(8069)應會繼續下跌至$15.1才反彈. 由於同仁堂是一隻優質股, 所以以適宜在此價入貨.SCMP的反彈似乎己告一段落, 下周有可能跌落 $2.85

星期二, 3月 21, 2006

同仁堂(8069.HK) 為何下跌

最近發現 同仁堂(8069.HK)走勢很奇怪, 從高點$17. 3, 一直往下跌. 但同時, 基金卻一直在默默吸納. 見下表, 而且大部份是在場內買入. 不知是否內有玄機.

星期一, 3月 20, 2006

Next Media will drop

Here some rumors that the competition of Chinese Newspaper market will intensified. More free newspaper will enter the market to share the limited advertising revenue. Base on the daily forecast line above, Next Media will drop to $4.12 next week. I sold the stock on hand with 30% profit and hope to buy it back below $4 shortly.

星期六, 3月 18, 2006

Forecast on SCMP (583) using Stepwise AutoRegressive algorithm

WeiQiao Textile raised around 4% on Friday. Based on the recent trend, it keep floating in the range of $11 and $13. Besides, by using the expontential regression, I expect this stock will drop back to below $12. In order to realized the profit, I sold a lot of my holding yesterday, which I bought at $10 per share, with 24% profit margin. I am expecting to buy it back at the price below $11.5 in next 2 weeks.
Recently another stock, South China Morning Post (0583.HK), catch my attention. This stock perfectly fit the requirement for a good stocks according to Warren Buffett's stock picking strategy, SCMP fit the following factors.

1. It has high average Return of Equity and Return of Asset ratio (consistantly above 12%)
2. It has low Debt/Equity Ratio, that means low pressure on interest increase.
3. It has a almost monopoly status in HK's English newspaper market.

Recently SCMP is suffering serveral unfavorable factors which cause the stock price to dropped quite a bit, including,

1. The listed HK stocks need not to put their announcement on the newspaper, that will cause 5%-10% reduction in the advertising revenue.
2. Competition on the classified post, from both other English news paper and Internet job Posting Sites.
3. Another possibe factor maybe the shrinking of the Engish Readers in HK after the decrease in influence of British government.

I estimate the buying point using SAS Forecast, this time instead of using Expotential Smoothing, which I tested with the history data, is less accurate that another methods, Stepwise AutoRegressive algorithm and have the following result.


The SCMP is expected to get back to around $2.9 and remain flat. I think it will be a right price to buy the stock at around 2.75, which is also the stock buy back price by the SCMP last year.

Maybe Fine tune needed in my forecast

Cheung Kung (0001.HK) will annouce final result next week. According to the record that Li Ka Shing kept buying the stock, it is very likely that the result will be a good one. That make me worry that next week it's stock price will raise instead of drop as forecast in my regression chart. Maybe I should use the daily data instead of weekly data to reforecast the trend of the stock.

星期五, 3月 17, 2006

魏橋紡織可能下跌

魏橋紡織一向是我的幸運股, 幾次買進的貨都有利潤. 但此股的波動性很大, 令人很難放心長期持有. 但從另一方面考慮, 它卻是短炒的良品. 今天魏橋紡織突然發力從$11.95衝上$12.45. 這次我跟據魏橋紡織自上市而來的每日數據, 用Expotential Regression的方法得出以下結果.其於以上預測, 下星期魏橋紡織有可能跌回$12,更可能跌至$11.7. 為支持自己的測試, 我己在今天在$12.4放了一些持股, 希望SAS的預測不會讓我失望.

星期四, 3月 16, 2006

長實仍然在下降軌

長實(0001.hk)最近在李嘉誠多次大手買入上升後又再回跌. 相信誠的眼光, 今天用預測軟件計一下下周走勢. 數據是長實過去三年每周股價. 下周再看準確度如何.
從上圖, 長實仍然在下降軌, 下星期將跌致$78.5.